AI in organisations: what works and what's changing

Dawn Mischewski: All right, so welcome everybody, to this session on AI transformation in organizations. So, today's session is a joint presentation from Superesque and Helen Hovenga. It brings together, the Superesque community, and the HREC community And friends and followers of both. So welcome, everybody. It's great to see so many friendly faces. For those of you who don't know Superesque, Tim and I co-founded Superesque in the second half of last year. And we have spent the last year teaching non-technical professionals of all stripes how to think with AI and use AI to augment their expertise. Helen was one of our, I think, Cohort 4 students in Foundations. She has probably done the Superesque courses more times than anybody else. She attended Foundations several times, and she's now onto her third round of Pro. So, Helen, I think you take the prize there. Tim, for those of you who don't know him, is a philosopher by background, an educator by, I've lost the word… temperament.

Tim Rayner: Trade.

Dawn Mischewski: And trade. And he also gives a good talk. So, I am not going to spend a lot of time, introducing. I just want to do a little bit of housekeeping before we start. So, both Tim and Helen have assured me that they will speak for no longer than 20 minutes each. Hopefully that will allow us 15 minutes for questions at the end. Put your questions in the chat. I will consolidate them as we go. We'll do a round of, kind of, anonymized questions in that last 15 minutes. And then, when we hit 1 o'clock, we will turn the recording off, and if anybody wants to stay and have more open conversation, that is fine. That's the slide for the end of the talk. With that, I'm going to hand over to Tim.

Tim Rayner: Thank you very much, Dawn, and great to see so many friendly faces here. Lots of Superesque students from over the cohorts, but also a bunch of new people, which is really exciting. So, let me just bring up my slides. And I will get started. Okay, Dawn, is that looking great?

Dawn Mischewski: You've still got your band across the top. Perfect.

Tim Rayner: Oh, sorry. Is that… is that right, or has still got the band?

Dawn Mischewski: Now you've got the band back.

Tim Rayner: How's that?

Dawn Mischewski: Yes.

Tim Rayner: Perfect. Okay. Now, Dawn, mute yourself, because I can hear a bit of echo. AI has only been with us a handful of years, but it's already having significant, impacts on our organizations. It's… there's no doubt that it is, changing the way that we work. And leading us to rethink the way that we should design, our work, also train for people and manage teams, there's a lot to think about. One thing we probably don't think enough about is how AI is changing us. And for all the HR and people and culture people in the audience, I'm talking specifically here about the people in your organizations. The question that I want to explore here is. how AI is amplifying human potential within the organizational context, and what people who are managing teams and managing people need to be thinking about in order to capture the opportunity of these changes. Right now, if you're working with AI and thinking about AI in organizations, you're probably struggling with a major conundrum, and we see this conundrum around the world. Australians, overall, are fairly heavy users of AI, But organizations are not yet seeing the productivity gains that we all anticipated. Research shows that somewhere around 59% of Australian workers use AI at work. Maybe not all the time, but they have used AI at work. That's up from last year. It's slightly down from the global average, which is around 64%, but it does indicate that Australians are really taking to AI at work. Another really valuable statistic here concerns the widespread shadow AI phenomenon. As I was saying, Australian employees are big users of AI. They're probably feeling a bit constrained within the enterprise AI environment. They're exploring AI outside of those environments to try and keep up. And another stat that bears this out is the fact that Anthropic's research shows that Australia is the 7th highest user of, Claude, per capita. It's just below the US and Canada. So, there's a healthy use of AI out there, but at the same time. Australian executives are not seeing that gain. 82% of Aussie executives say their firm has made… AI has made no difference to their firm's productivity over the last 3 years, and that pretty much lines up with the global average. So, that's really the conundrum that we're all facing at the moment, and we're all puzzling about why we're not seeing that productivity gain. I think there is a very common answer, and it's probably right. By and large, our organizations are not designed for AI, so what happens is that when individuals use AI and speed up their own work, it just creates a bottleneck further down the line. And even the big AI companies see this. Anthropic says that human review has become the number one bottleneck in their coding activities. So this is definitely one answer as to why we're not seeing that AI productivity gain, but I think the challenge is ultimately bigger than this bottleneck itself, what we need to reckon with is the fact that Our organizations need to be redesigned, in… from the ground up, essentially, for the age of AI. And this is something that everyone, all the leaders in this space need to be thinking about. We need to rethink the way we work, the way that we train people, the way that we manage people, and even the business models of our companies to unlock the productivity gain of AI. Now, history actually bears this out, I'm not just making it up. If we look back to the 19th century, there was a… three or four decade lag between the introduction of electric motors into factories and the point where industry actually started seeing productivity gains from electricities. Electricity, rather. It's not hard to understand why this was. Back in the 19th century, factories were driven by steam, all the machines were hooked up to belts that ran along a line shaft, and the steam engine turned a line shaft. That worked perfectly well, and so factory owners thought, well, if the line shafts work, electricity is still kind of expensive, why change the way we work? And indeed, it makes sense. By 1899, some 15 years after the introduction of electric motors. electric motors still supplied under 5% of factory power. So, what shifted? Well, a threshold moment, a big pivotal moment, came with the rearrangement of the Ford production line. You probably all know the story. Henry Ford decided that now we have these new lightweight electric motors. We don't need to run everything from these line shafts, we can actually give every team their own electric motor and rearrange the factory so the flow of work is actually much more efficient. And this led to real productivity gains. Between 1913 and 1914, the time it took to build a Model T chassis fell from 12 and a half hours to one and a half hours. But here's the thing. In the same period, workers just went apoplectic. They were being forced to work a different way, they didn't know what they were doing, it was a break from the usual flow, and so staff turnover spiked, it went through the roof. Ford responded by doubling pay, raising it up to $5 a day, and, interestingly enough, setting up an employment department to manage the workforce. Many commentators say that that was actually the origin moment of the human resources departments that evolved through the 20th century. So even that moment did not lead to a a wholesale transformation of industry. By 1919, electric motors supplied half of the factory power. It was really only once other industry leaders saw how rearranging their factories led to this productivity unlock that they jumped on board, and of course, everything changed. The rest is history. Now, this phenomenon, which you've probably heard of, is called the productivity J-curve. Productivity payoff of new technologies usually doesn't arrive until organizations redesign work, train for new skills, and implement new modes of management. We saw the same thing in the computer era. Back in 1987, Robert Solow, a famous U.S. economist, said, look, you can see computers everywhere, except for in the productivity stats. And that is because, at that point, companies were still fumbling around in the dark, trying to figure out how to unlock value from these new tools. We were in that kind of productivity dip. Computers were actually making work less efficient and creating complications. It wasn't until 10 years later that computing began to unlock its productivity gains, and those changes came with the emergence of a whole new set of ways of working and ways of managing Distributed decision-making, collaboration, agile teams, all of that led to the productivity unlock in the age of computers. So, I think that right now, with AI, we're in that dip of the J-curve. Individuals are seeing gains, but organizations aren't, because they haven't yet, or they are in the process of redesigning work and exploring new training and management processes. We're just at the beginning of that process. So, change is happening, I think. It's happening everywhere, but it's not yet showing up in the productivity stats, and I think the question that should be on every, HR leader's, or people and culture leader's mind is. what's our role, right? What role does AI… HR need to play in this process of redesigning our organizations? So I have a thesis, and hopefully you will find it enlightening. I think that in order to understand how to rethink the way we manage people and train people and enable people to be more productive and effective in the workplace, we need to understand what AI is doing to us. Now, this is interesting. Every technological revolution amplifies human capability, but it amplifies human capability in different ways. Electricity amplified workers' ability to move and make things. Computers amplified people's ability to manipulate information. AI, by contrast… well, I mean, AI does improve our ability to manipulate information, and if we're using AI agents, it amplifies our ability to make and move things as well. But what is unique about AI is that it operates on our intelligence and our cognition. So I think what AI is doing to us is that it's amplifying our power of thought, judgment, and expertise. And what HR and people and culture leaders need to do is to think about how to set up new training and management practices that can capture that increased Cognitive capability. Now, this is really, worth some serious thinking, because When AI amplifies our thinking, it effectively amplifies not just what we can do, computers amplify what we can do, AI amplifies who we are. Our ability to think and to act on our thinking is closely tied to our sense of personal identity. So think about it, you know, your computer enabled you to execute faster and come up with some ideas to a certain point, right? But AI amplifies your ability to see more. And understand more of what you see. AI gives you new skills that enables you to act on these new things you see, and it thereby increases your capabilities and your ambition. Right? So AI is actually amplifying people and amplifying their ambition to go out there and do and achieve. So I think that this is fundamental. to the challenge that lies at the feet of HR professionals today. Our challenge is to figure out how to manage employee amplification. And I see there are 3 jobs here, and I'll be quick about them because I don't want to cut into Helen's part of the talk. First of all, we need to figure out how to train people to really work collaboratively and effectively with AI, and I'll say a bit more about this in a moment. Secondly, and very importantly, we need to make people safe, to feel that they can share their tacit expertise. Basically, that means the real skill, the deep skill that they bring to work that makes them valuable. We need to make people safe to feel that they can share that. Once we've done that, then we need to think about how we can redesign roles and career pathways, so that these people, your best people, who are just getting right into AI and leveraging and expanding their capabilities and their ambition, don't feel trapped in their roles, and don't just kind of move on somewhere else, because I think there is a battle for talent coming up in the age of AI, and, you know, your organization wants to win that battle. You don't want to be the organization that's just kind of bleeding good people, because they are stuck in roles, and they cannot evolve fast enough. So let me just go through these three, challenges, and three jobs, and I'll do it quickly because I want to stay, In my time. How much time do I have, Dawn?

Dawn Mischewski: I think you're on track. You're at 12.20, so you've got at least another 5 minutes.

Tim Rayner: 5 minutes. Okay, well, let's just step through these three jobs, because I think this is probably the most important, most relevant part of my talk for the audience. So, starting with this first job, training people to work collaboratively with AI, this is something that we do at Superesque, and so we've learned quite a lot about it. And the key thing that we've learned is that to collaborate with AI and get the greatest value out of that, we need to… surface and articulate key elements of our own deep expertise and share it with the AI, so the AI is working within our frame of reference and can actually effectively collaborate with us as a teammate to get things done. Doing this is not an easy task. We're talking about surfacing tacit knowledge, and it's well known that tacit knowledge is deeply buried in our bodies and minds, so it takes a good bit of work and skill and training to get that stuff out. But I think this is the job that HR is faced with today. We need to learn how to train employees to surface their tacit expertise, that's the knowledge and expertise that they've accumulated over years, that they use every day, that enables them to get excellent work done. And we need to find a way to convince them to encode it, or share it, encode it into the AI systems they work with, so that together, the human and AI partnership can be more than the sum of its parts. So what we need there is real AI education, not just. ChatGPT training or… or prompt workshops. Now, of course, you're already thinking, this is a significant challenge, and you probably already see the challenge coming up. people, in general, will feel very uncomfortable about surfacing their tacit expertise and encoding it into AI. No one wants to feel like they are building the next generation of AI systems that are going to replace them in the workforce. I can offer a bit of advice here, and this is based on what we've seen in Superesque. We teach people to build AI agents trained on their expertise, and what we've learned is that even the best trained agent is no substitute for for the human expert working with the agent. I mean, really, we need to train people to build AI agents and assistants that they can work with collaboratively, and then we need to incentivize that whole process, and figure out a new way of managing human and AI teams that really raise people up and enable them to do their best, and beyond their best work within the workplace. So we need to be thinking about new incentives to incentivize that surfacing and encoding of tacit expertise into AI systems. But we also need to be thinking about how to recognize people for doing that, because we don't… we don't want people to be thinking, oh, great, I share my expertise with AI, and my employee goes, hey, thanks very much for giving us your secret sauce. You're… you are… you're relieved of your position now. Imagine, instead, an organization that said, look. Thanks for being our most valued AI-enabled expert. We're building a culture of AI-enabled experts, and you're one of the best. I mean, I think that is the kind of culture that organizations need to be, focused on creating at this point. Finally, once you've learned how to amplify people and make them safe. to work with you to achieve that, we can all settle into the job of figuring out how to redesign roles and career pathways that fit into the new emerging structure of our AI-native organizations. It's very important that people and culture think about these new career pathways because, as I've mentioned. and we see this in Superesque all the time, people who know how to build AI agents trained on their expertise, they have superpowers. In fact, one of our participants in Pro today said, we're like gods! Which, it was a bit ironic, but there's a certain truth in it. People who are doing this work will quickly outgrow the roles that you designed for them prior to the AI age, and if there's no room to grow. you're gonna lose your best people. So, it's very important that HR, people, and culture thinks about how to redesign roles and career pathways that grow with people, so people not only feel enabled to build these human-AI partnerships and feel rewarded within the workplace for doing that, they know that by doing that. that's going to take them to new and exciting positions within the organization. So just to sum up, I think my basic point here, my key point, is that HR emerged with the 20th century organization, but people and culture is at a point where it needs to reinvent itself, for the 21st. There's a lot of work there. If you need help, we've actually launched a free program just today called Superesque Start. You can access that for free from our website. Takes about 4 hours to do. It's for people who are using ChatGPT Claude and Copilot, and you can complete the program on your own system using whatever AI tool you want to use. So, hopefully that helps, and hopefully the insights I've shared in this talk have also helped your thinking going forward. Thanks very much.

Dawn Mischewski: Thank you, Tim. Helen, over to you.

Helen Hovenga: Hi, everyone! It's, so good to see so many familiar faces. I'm looking forward to sharing some, some thoughts with you today. Hopefully, can everyone hear me okay? Yep, all good. Beautiful. All right, off we go. So, Great to hear, Tim's introduction, and it's been, certainly a great collaboration working with Superesque on this. I guess a quick introduction to me. My name is Helen Hovenga. I'm excited to share my journey with you today. I've been in people and culture functions, for 25 years. And working with many different and great organizations, who I've learned a lot from, and 10 years at the executive table, called the CPO, or a General Manager of HR, or an Executive General Manager of People and Culture. For the last 4 years, I've been working as an interim chief people officer, so I'll talk about that in a moment. And I guess what is key is that I love AI, and the reason that I love it so much is because it is the intersection between two things that I'm passionate about. One is technology, and the other is people. I also like AI, because it makes my brain hurt. It makes me think more deeply. So, in having gone through the Superesque course. Instead of thinking, I'm learning about a technology system, I've now understood that I'm opening a door to understanding myself better, and being able to enable myself to broaden from that expertise as well. I thought I'd briefly show you an example of, you know, the last 4 months of work. It's certainly slowed down now, which is good, but it was really exciting, you know, from April to September to be working with a number of different organizations. At the moment, I'm working with only EPSA, but I was working with a number of them, and this was a typical week, so I'm showing you an example of both being an interim, which means that I'm there for a short amount of time, sometimes 3 months, sometimes that extends to 12 months in organisations. And for somewhere like Nuchev, I was there for 3 years, but only working 1-2 days a week. So, and then, you know, Impact Global Health was about one day a week, mostly in the evenings, because it was a global organization, and then working with Make-A-Wish was about, four months and about half a day a week. So, being able to fit in all of this work, AI certainly enabled me to be able to, switch between them very effectively. I worked with many different tools in doing that, so Cowork. Cowork that had a Claude plugin, Claude, and of course, Copilot. TypingMind was what I was introduced to in, in, working with Superesque as well, and TypingMind is where I learned to create agents. So during this time, I was really on a mission to advance my AI capability, and I really wanted to feel empowered in using AI. So I did that through the Superesque Training Program, and it's one of the reasons that I was so enthusiastic about doing it several times over, so that I could deeply understand it and apply it. I do feel like the learning I've had with Superesque has made me to feel really confident across any large language model at all. Any LLM is fine. I'm very comfortable with all of them. Life is a bit less hectic at the moment, and I'm able to focus a little bit more on, sort of, developing my own business. So, it's been an interesting experience. My mission, though, and this is what I really wanted to get to, and it certainly is great to hear Tim talk about the imperative that we have in front of us, it's to empower chief HR officers to lead the AI transformations in their organisations. And I'm keen to talk to you about a bit of the journey there, and how that might be the case. So why, why the Chief HR Officer? An IBM study in 2026 of CHRO stated that employee confidence in AI decisions is rising in 63% of organisations where the chief HR officer shares ownership of what stays human. So without them, it was 19%. It also matches Stanford's research on 51 successful AI projects. This is quite interesting to see their enterprise AI playbook. They found that 33% of the hardest problems were people problems, so getting people to change and adopt was really the hardest part. However, 46% of organisations leave the Chief HR Officer out when the AI strategy is defined. So, AI adoption is about redesigning work and addressing the people problems. It's the CPO's bread and butter that we do that. So if you can… if you can, I think leading this within your organization, working hand-in-hand with the CTO, Chief Technology. they certainly bring the system, they understand the know-how, and all of the, cybersecurity aspects that are also important, the data aspects, and some of the development. But you bring the… you bring the people, and you bring the design, work, and able to build the capability as well. There's one other reason that I was wanting to, talk about why Chief HR, And it's because we understand what it means to hire a teammate, to work with a teammate, and to be able to get them on board and set them up for success and manage their performance as well. So, if you're doing that with an AI teammate or an agent, so when I talk about an agent here, I'm talking about an agent that is like a role with a character, it has instructions, tools, and decision rights, to a degree, within parameters. And that's what I'm describing could, make it the most effective way to be contributing to you. So, having a position description for your agent is really critical. The RCICE model, is certainly very helpful in understanding and thinking through the way in which you can embed the agent in your work. Set them up for success by providing the knowledge files, the tools, the connectors, the decision rights, and then help them understand the safeguards, so they need to know what they mustn't do, as well as what's expected of them. And then, of course, in managing their performance, have you tested it? Are there KPIs? Is there a feedback loop? This is… this example here is just one of many parallels within our role. Even designing an organisational structure that defines the roles of people and the roles of agents and how they interact is certainly something we can do very effectively as well. So I thought I'd start with my learning journey, just briefly showing, where I have started. So, I've been playing with AI for quite a while. As soon as it sort of came out, I tend to be an early adopter of anything technology-related, but it's really been since April that I turbo-boosted my work with AI in both my work and my learning. And again, the Superesque program was the whole world of introduction and feeling more empowered. And certainly enabled in using AI effectively. So, I was so enthusiastic in everything that I learned that I took it back to my people and culture team and said, hey, I've got something new I want to share with you. And so I did that in fortnightly meetings with them, and so they went on the learning journey with me, which was fantastic. It also embedded my learning better, of course, that, that I was able to share that. The systems I use, so I use Notion and TypingMind in the Superesque course, and then in Superesque Pro, I've been learning to work with Claude Code, which is, blowing my mind every day about what is, what is capable of doing. All of the training for my people and culture team, though, I've had to transfer the learnings that I've had from Superesque into Copilot, and now to Co-Work as well. But of course, I'm learning as I'm going and using the systems and delivering as I go as well. There were a few surprises in my learnings in the journey with Superesque. The first, and the most surprising, was the fact that the activities that I had to do. So in the first week, I thought, great, I'm about to learn how to make an agent, finally. And instead, I spent a week learning about myself, how I think. And my tacit experience that, that Tim has spoken about today. The second surprise was that, the agents who knew my way of thinking and my knowledge and my charter and my ways of working, they're the ones that have made the biggest impact. So, in being able to embed my tacit knowledge into those agents, I think that's made a big difference as well. And then the third is, that every agent needs such significant testing. It probably goes without saying, but knowing that if you're going to launch an agent out there, the degree of testing that's needed, I hadn't anticipated that. I like the idea that I start something, get it going, but all that testing stuff. Not so keen on, but the reality is, of course, the outcomes are fantastic when we do that. So, Cowork tests an agent automatically, which is helpful. So, you know, some call it skills agents, but essentially, it's, it's testing it as it goes. But I do believe that you still need the learning loop to gain feedback for it to be most effective, so I have found that Feedback is one of the key aspects of success in working with any agent. The other thing I thought I'd share was this, the structure of my, of my agents and the business that I, I have and I'm building. So on the left-hand side, that's my team. So that's what's running on Claude Code. Some of these agents were designed by Superesque and provided to me, and others were a part of the onboarding of agents when I became a part of the Superesque Pro course. So, Aria is my Chief of Staff. I call her Aria because she makes me sing, makes my life a lot easier every day. Even today, she said to me, oh, you're going to be out at the time that you're presenting, make sure that you've got the right environment to be able to present in. So it was nice to know that she was thinking of me. Coco's my thinking partner, so that's the essence of my own thinking. And COCO stands for Chief Organization and Culture Officer, which may be the name for us in the future. Beacon helps me with written pieces, so it might be a LinkedIn article or a blog. The majority of that I write myself, but I certainly like Beacon's support and help with that. Sextant does my research. Loom is my designer for agents and other capability solutions. Forge is a builder of systems, websites. ATLAS is a librarian, so it manages the knowledge base and the information that's being created. Crucible is the evaluator and keeps assessing every agent so that every day it's getting better. And on the right, there's some of the agents I can build for organizations. So, they can be developed on Claude Code, and then they'll be placed in whatever systems the organization uses. So, if there are any organizations that are needing me to design or develop agents. then I can tailor that to what the business needs, embedding that with the company values and context, so that it understands the work that it's doing. And underneath all of that, of course, is my purpose, which is to leave every person we work with more capable and their organization stronger for it. In being an interim, I love the idea of making myself redundant at the end of every project that I do, by building the capability within the organization. My vision is for a world of work where people and AI are true partners, and change is done with people, and not to them. And then, of course, my values. So my values include starting with trust. Courageous candor with kindness, human and grounded, and change that lasts. And excellence, always. I… I found a quote that I thought was beautiful, and I think, Annabel, I think you're on the call, and I think you posted it in one of our Slack chats. So, I'm not going to fall in love with robots, so don't worry about that. But I really do like this saying. It's from an article that was back in March 2026 by Barton Friedland. His doctoral research looked at how technology takes part in leadership and decision making, and his key point was to encourage people to invest in how people work with AI as much as in the AI you buy. And he said that success or failure does not reside in the intelligence of the model, nor the capability of the human, but in the quality of the relationship between them. So it certainly talks about synergy, what's the communication? What's the feedback, do they understand each other? And just as a relationship amongst humans matters. You can see that the relationship between the model and the human matters as well. for the People and Culture team training journey, I know there were some questions about what exactly would be part of the learning. So, this was the journey that I undertook with the People and Culture team when I met with them fortnightly on their, AI learning journey. So, certainly they can build their own skills and agents now, they apply that every day, and, and they're thinking through as they're doing their work, how can we maximise the use of AI? In the foundations, they developed their own context card, who am I, what do you already know about me, and making sure that that's part of their standing instructions. We worked together on designing agents in Copilot Studio, and then capturing their own expertise as individuals. And then we've worked on skills in Cowork, and worked on really creating better prompts, and, and the prompt library's been created as well. We talked a lot about tokens, and how to make the most of the work that they do, and making sure that they haven't, sort of, created AI slop. So, one of the things my team knows is that I'm not as interested in looking at the output until I know what the input has been. So the questions that I ask my own team before they show me anything, or if they've sent me anything, the questions I ask are, what research did you undertake? what prompt did you use, what's the thinking you've put into this, and what consultation has gone into this before you've then created this output? And once I know that they've done the appropriate thinking, then I'm more than happy to look at the document that's been created. I know that also some of you would be interested to understand what would be possible in the People and Culture team, and I saw a question that said. a lot of people are doing work in, talent acquisition at the moment, because it's one of the easiest to say, to apply AI to, to automate a lot of things. I think there's going to be quite an important decision on privacy coming up that we need to be aware of and cautious about in talent acquisition activity, so do, look it up, it's starting in December, but, talent acquisition's going to be one area where we'll have to have more transparency around the decisions that have been made about people's futures. Because decisions are being made by robots. So, let's be aware of that. However, talent acquisition is kind of an easy, low-hanging fruit, and I encourage that work to be done. In fact, it's one of the first agents we have been creating at EPSA. In, in other opportunities, though, in having, as a Chief People Officer, I have Coco. There's also a Chief of Staff, obviously I have my own, and then an OD Communications and Change Advisor would also be possible. Under the business partners, you can see, there's a few different options. So, I just wanted to point out that, it… the benefit is actually in having The humans have agents that they partner with, and that ensures that there's more time for the people interactions and the genuine connection. When I first came to the team, they were frustrated by the administration that they needed to focus on. These are not all in place within EPSA, but these are potentially what could happen in the organization, and certainly the structure of the people and culture team. And you can also see that there's a foundation for all of this, and that is the character. So what's the temperament we want a people and culture agent to represent? What values? What voice? What's the team constitution? What are the ways of working in organisational context that would be critical? So all of those elements would be standing instructions in anything that, that is created. as well. You may have some other questions about that, but, let's, let's see how we go. So, in the chat, I'd love to know who's accountable for AI in your organization. We'll certainly be taking note of, of what's happening. It may be nobody yet, that's okay. I'm just interested to understand Where that is. And one of the reasons that I ask that is because 36% of executives say unclear accountability complicates AI deployment, and I've certainly seen that firsthand as well. Where we're waiting for someone to take charge. There will be lots of people running in lots of different directions, but they may not be aligned in the direction that's needed. So, AI transformation requires the ownership of leaders and people, and I really strongly believe you need to know it to lead it. This is not one where you can delegate this. This is something that, if you are leading it. you do need to know it. So if you're looking to analyse or lead an AI transformation in your organization, here are the elements for a successful organization-wide deployment. So I've developed these based on a few different resources. The Stanford's Enterprise AI Playbook, that was taken from 51 successful AI implementations. And then, Professor John Whittle, has also done some work on this, and then I've added some from my own, first-hand experience and observation, as well as 25 years of change management within my, people and culture function. So, it came down to four different themes, leading it from the top, and ensuring that there's ultimate accountability. The funding and the proof of the value, and this is where we're in the J-curve at the moment, the value's not necessarily there yet, but we need to start to focus on value-add projects with AI. Bringing people with you, and then governance. The QR code, that provides a link to doing a self-check, so if you want to get your phone and save that, have a look. There's 27 elements and 4 different aspects, and it's also available on my website if you want to come back and have a look at that later. So the fundamentals I would point to, if you did nothing else of the 27 elements. Probably the executive sponsorship, a people and culture leader leading it, having a culture that encourages AI use. Building capability, and then providing strong governance that is there to speed up deployment rather than slow it down, I think are all the critical aspects that I'd ensure are a part of your AI transformation. Also, as we know, in any change management, we… there are many concerns, and I speak to a range of people, and hear all sorts of concerns, and I appreciate that people are really honest with me about these. I don't think they can be ignored. So, probably number one, the most significant one, is that employees really fear for what AI means for their jobs. My position on is that AI changes jobs more than it removes them, and I think that's important to acknowledge that. If we can anticipate the changes before they start to have impacts, then I think we can really help people with this. So AI will redesign roles. You simply look at the people and culture structure that I provided earlier, and you can see how, you know, this will really change the work being done by people in the team. This will increase their interaction. There'll be more time for thought, more time for people, interaction, and less administration as well. So the Stanford… Stanford, research suggested that in those successful projects, so, the 55 successful projects that they researched, headcount fell in 45% of the successful projects, so it did have a… it did have an impact. But in another 55% of the organizations, they… some of them avoided hiring. Some of them, cut nobody, and, and some of them moved people into other work. So, Anthropic have also published a recent article where junior technical hires fell from 27% to 9% since 2025. So, that is having an impact on our junior employees. interesting to see Anthropic's studying themselves and seeing the impact of AI on their own organization. I think on this one, the fear is understandable and real, and there are impacts that we need to be aware of, but also anticipate. I do believe that, with junior employees, we have one of the best opportunities, if they're early adopters, to sort of take on the magic of AI, ensure that they're working in partnership with those who have the wisdom and experience of many, many years, and as a joint force, I think that can be very, very powerful. Bringing together your junior employees with your more experienced, wise heads, I think, is one of the things we can make a difference with. There's probably 4 anticipated challenges, and I'm sure there's many more, that really come to mind for me. The biggest question for me, and I really don't know the answer to this one, is who owns what I have built? Tim talked about this. Putting one's tacit knowledge into… and your personality into an agent. What happens to that agent when we leave? Is this transferable? So if I leave an organization. Do I also have to rebuild the skills that I've been building in the systems that I've been using in every role that I move to? They took time to develop, and I think that's one of the challenges that we have? Do we then have to have new employees add a whole lot of time in setting up and developing those skills again? So the agent ownership might be something that needs to be added to a contract. You know, where there are skills, I think it's going to be important to understand where they sit. The second one, then, is tokenomics. So, tokens, the budget line, I think businesses are really starting to see the cost, and it's mounting up. It's very easy to waste tokens. My team also know that in doing their training and having some standing instructions, they are far more efficient in their token usage In the ways in which they're working with, particularly Cowork and co-pilot, because of the training that they have and the capability. So that's the key answer, I think, to maximising token spend. The third one, then, is facing ethics deliberately. So what's our position on jobs? The environment? The cost? the privacy consultation, what's your policy? What's your strategy, and do you know where you stand as an organization? So, in each of the organizations I've worked in, one of the first aspects has been ensuring that they have an AI policy that actually has The balance between the security aspect, but the freedom to be able to try things as well. So, One of the last is Shadow AI. Tim also mentioned Shadow AI use. This is where people are really potentially taking company data and using it in their personal LLMs at home, and they're doing that sometimes because they need to get the work done, but they can't see how the systems that they have at work are going to help them to get that done. And then they don't tell anyone about that. So some organizations have been so conservative, and they've been able to prevent it, but there are many who are doing it. So it's important that those people, that are… we define some lanes and tools, understand what their challenges are, train them in ethics and awareness of where the data goes and the implications of that. So I think helping them understand what they can go ahead with, what they need to treat with care, and what they need to stop doing is also important. And having said all of that, what I love about the Shadow AI is that they're at least having a go, they're trying it out. It might be, causing mistakes at times, but, I think if they do that with caution, we need to be able to bring them into the fold and make them champions. I'm almost there, I think we're getting close, aren't we? Probably over time. Case studies, so quickly, EPSA, I guess, we have some of the elements in place. We don't have all of the 27 done yet. We're certainly working on our capability building. I've trained all of our senior managers. We're starting to recognise the need to get expert knowledge into AI, and we're early on the journey, but steps are at least in place. The other case study I thought I'd show you is Make-A-Wish. So in working with them, they asked, you know, how do we bring AI into an organization whose whole purpose is human? The mission for them is that every dollar and every hour goes further to grant more wishes, of course, for more children. So I started with, sort of, a learning experience, asking them what systems they're using, what took their time, what was their attitude to AI, I think everyone felt really heard. It brought people, to open their minds towards AI and become more curious. I presented a 90-day plan. Those actions have now been put in place as well. And then, just on the bottom, I have a plug, in speaking to the CEO at Make-A-Wish. They've got Trees of Joy. It's a fantastic team-building opportunity where you get together and decorate Christmas trees. I know many of you might love that. And if you're interested, they've opened it up to us So scan to register, and take your team. It's teams of 10, and you can go and build a wish… a Christmas tree for a wish child. Last thing, my principles, so match to the elements. Again, very quickly. You need to know AI to lead AI. Stay 6 months ahead. The executive team are going to ask questions. Be aware of it. I have a healthy dissatisfaction with my level of awareness in this space. I want to continue to learn it. Help people through change. AI amplifies what is there. If you've got poor data, you'll get poor outcomes. Give it a try, and then input before output as well. to close out, I just wanted to, let people know, sort of, the range of work that I've done in my interim and fractional work. And then, how I can also help. You can see that I work from organization-wide, down to a people and culture team, and then individuals. So at the moment, I'm coaching one-on-one, a person who's building her AI-native, business as well. So, I guess we can go on to questions, as well. Interested to know what's going on. I did want to let you know that, I'd love some feedback at the end of this. I'm like an AI bot. I love the feedback. It makes me better every day. And also, if you want to scan for my website. I'm setting up some time on Fridays from 2pm. If you take the link there and just join on Zoom, we can all have a chat together. about AI. That's it from me, on to questions.

Dawn Mischewski: So, Helen, I have two questions that have come for you through the chat. First one, I'm scrolling wildly for it. Okay, so the first one comes, and I'm also very curious in the answer to this one, how long did you spend building your organization, your AI team?

Helen Hovenga: I'm gonna end my show so that I can look at you, if that's okay? It's a… it's… you know what? It's actually a really quick thing. If you can imagine how busy I was, I've lost you all, hang on, just a tick. So… if… if I talk about building an agent, I'd say, certainly within a day of work, if you add it all together, I'd have a pretty comprehensive agent done. But it's all in the refinement. So if I think about the TA agent, that's taken a long time, because I've wanted to consult with people, see how it's working, test it out. We're using it as we test it, so that's working well. But, you know, that's certainly taken a bit longer, because it's also a team effort, that one. But the ones for myself, it takes no time to develop agents. Once you know what you're doing, it's very easy. For the team of agents, what I started with As a typical CPO would do is an organization structure. So I worked out, what do I really need here? How do they fit together? What are the big pieces of work? Just used an operating model, I guess. what do I need to deliver, what are the pieces of work, and then what are the aspects of that? And then that defined the sort of roles that I needed, working out what's human and what's an agent. I hope that helps.

Dawn Mischewski: Thank you. I'm just going to pause. We have a couple more questions, and I can see some hands up. I just want to pause for anybody who needs to leave at 1 o'clock. Just one quick reminder, we will be sending recording transcript slides. later on this afternoon. So, thank you for everybody who joined. We will keep talking, but, it's been a great turnout. We're really pleased, and thank you both Tim and Helen, for the contribution. But having said that, Helen, I have another question for you. So, have you faced a situation where the person an agent reports to leaves? How do you handle the reporting line so that important work the agent is doing doesn't get lost?

Helen Hovenga: I haven't… I haven't had that, but, I do, you know, it will be a really critical part of your thinking and planning prior to creating an agent, is ensuring that it is trans… it's easily transferable, when it comes to the ownership. So, plan it ahead. Plan ahead. And ensure that there is a handover.

Dawn Mischewski: Okay, so third one, and this is a comment from a Superesque Pro graduate, who says. Where has it gone? Can you make sure that somebody mentions that Cowork is a ball breaker and makes you go backwards? So for anybody who's working with Cowork, you can take that or leave it. And then we have one more in the chat. With AI agents, how do we assess and manage the security risks, and how do we introduce agents into an environment where governance and policies haven't yet been fully established within the organization?

Helen Hovenga: Yeah, I…

Dawn Mischewski: In this one, I'm turning the recording off and opening it up.

Helen Hovenga: Yeah, I, I wouldn't… I wouldn't unleash an agent on an organization without guardrails. So, quickly write a policy, and then you can launch your agent. But do the appropriate testing. You know, I think it needs to be practical. I don't want to hold people back, but it's pretty important that there's some guardrails in place.